Fluorescence-guided surgery is set to play a pivotal role in the intraoperative management of pediatric tumors. Shortwave infrared imaging (SWIR) has advantages over conventional near-infrared I (NIR-I) imaging with reduced tissue scattering and autofluorescence. Here, two NIR-I dyes (IRDye800CW and IR12), with long tails emitting in the SWIR range, were conjugated with a clinical-grade anti-GD2 monoclonal antibody (dinutuximab-beta) to compare NIR-I and SWIR imaging for neuroblastoma surgery. A first-of-its-kind multispectral NIR-I/SWIR fluorescence imaging device was constructed to allow an objective comparison between the two imaging windows. Conjugates were first characterized in vitro. Tissue-mimicking phantoms, imaging specimens of known geometric and material composition, were used to assess the sensitivity and depth penetration of the NIR-I/SWIR device, showing a minimum detectable volume of ∼0.9 mm3 and depth penetration up to 3 mm. In vivo, fluorescence imaging using the NIR-I/SWIR device showed a high tumor-to-background ratio (TBR) for both dyes, with anti–GD2-IR800 being significantly brighter than anti–GD2-IR12. Crucially, the system enabled higher TBR at SWIR wavelengths than at NIR-I wavelengths, verifying SWIR imaging enables high-contrast delineation of tumor margins. This work demonstrates that by combining the high specificity of anti-GD2 antibodies with the availability and translatability of existing NIR-I dyes, along with the advantages of SWIR in terms of depth and tumor signal-to-background ratio, GD2-targeted NIR-I/SWIR-guided surgery could improve the treatment of patients with neuroblastoma, warranting investigation in future clinical trials. Significance: Multispectral near-infrared I/shortwave infrared fluorescence imaging is a versatile system enabling high tumor-to-background signal for safer and more complete resection of pediatric tumors during surgery.

Fluorescence-guided surgery (FGS) is used in many pediatric subspecialties but there are currently no standard guidelines or outcome data. We aimed to assess the current status of FGS in pediatrics using the Idea, Development, Exploration, Assessment, and Long-term study (IDEAL) framework. Clinical papers on FGS in children published from January 2000 to December 2022 were systematically reviewed. The stage of research development was measured considering seven fields of application: biliary tree imaging, vascular perfusion for gastrointestinal procedures, lymphatic flow imaging, tumor resection, urogenital surgery, plastic surgery, and miscellaneous procedures. Fifty-nine articles were selected. For each field of application, the overall IDEAL stage was determined to be 2a for biliary tree imaging (10 publications, 102 cases), 1 for vascular perfusion for gastrointestinal procedures (8 publications, 28 cases), 1 for lymphatic flow imaging (12 publications, 33 cases), 2a for tumor resection (20 publications, 238 cases), 2a for urogenital surgery (9 publications, 197 cases), and 1-2a for plastic surgery (4 publications, 26 cases). One report did not belong to any categories. FGS in children is still in an early phase of adoption and development. We recommend using the IDEAL framework as a guide and suggest developing multicenter studies to define the standard guidelines, effectiveness, and outcomes.

Biological evaluation of anti–GD2-IR800 and anti–GD2-IR12 using a preclinical NIR-I imaging system in vivo. A, IVIS Spectrum images of the exposed tumors at 24, 48, 72, and 96 hours postinjection of anti–GD2-IR800 (n = 3) and anti–GD2-IR12 (n = 3) in mice bearing subcutaneous LAN-1 neuroblastoma tumors. White dotted lines show the ROIs drawn to quantify tumor to background ratio. B, Graph bar showing fluorescence intensity of the tumor and the background at 24, 48, 72, and 96 hours post conjugates injection. C, Tumor-to-background ratio at 24, 48, 72, and 96 hours after the injection of the conjugates. Error bars represent the SE across individuals. D,Ex vivo fluorescence images of the resected organs. The white dotted lines delineate the regions used to quantify the fluorescence signals. E, MFI for each organ at each time point and for the background card (BG). Fluorescence intensity decreases over time, with the anti–GD2-IR800 being consistently brighter than the anti–GD2-IR12 (MFIIR800/MFIIR12 = 1.90 ± 0.36 for tumor, P = 10−20, three-way ANOVA). Error bars represent the SE across individuals. F, MFI of tumor relative to MFI of each organ at each time point. Data for control mice is shown in green in the graphs B, C, E, and F.

In vitro multispectral NIR-I/SWIR fluorescence imaging of anti–GD2-IR800 and anti–GD2-IR12 A. Schematic representation of selected cell lines (left) and number of cells used (right) to assess the camera sensitivity. Cells were stained with 100 nmol/L of anti–GD2-IR800 and anti–GD2-IR12 and prepared as pellets in microcentrifuge tubes [tubes 1–4, GD2-positive; tubes 5–8, SUPT1-WT (GD2-negative)]. B, Multispectral NIR-I/SWIR images were acquired. Images are shown for 1,050-nm LP filter LAN-1 cells (tubes 1–4) and SUPT1- WT cells (tubes 5–8) with anti–GD2-IR800 and anti–GD2-IR12 on the same fluorescence intensity scale. Yellow dotted lines show the approximate positions of Eppendorf tubes. C, The SBR was calculated as the MFI of each cell pellet divided by the MFI of the control cell pellets for the corresponding cell line and dye. Data are shown for the 850-nm long-pass filter. Error bars are derived from the SEs within the signal and background ROIs. D, MFI of anti–GD2-IR800 and anti–GD2-IR12 decrease towards longer wavelengths. The shaded region represents SE over pixels in ROI. Exposure time = 25 ms.

In vivo validation of tumor uptake of anti–GD2-IR800 and anti–GD2-IR12 using a commercially available clinical NIR-I imaging device. A, Comparison of fluorescent images of the unexposed tumors at different time points postinjection of either anti–GD2-IR800 or anti–GD2-IR12. The laser boost setting of the EleVision IR Platform was set to arbitrary units for 24, 48, 72, 96 hours and exposed anti–GD2-IR12 images, respectively. B, Bar chart showing MFI of the unexposed tumors post anti–GD2-IR800 and anti–GD2-IR12 injections. Quantification was performed using the EleVision software. C, Bar chart of the TBR. Error bars are calculated from the SD across 5 points manually selected in each ROI. Anti–GD2-IR800, n = 2 at all time points; anti–GD2-IR12, n = 2 at 24 hours, n = 1 at 48, 72, and 96 hours. D, Images showing stages of tumor excision after injection of anti–GD2-IR800 under white light observation and NIR-I fluorescence. A residual tumor (3 × 5 mm) was identified on fluorescence imaging and subsequently excised. E, Histopathologic evaluation of both the main tumor and residual tissue, confirming the presence of viable neuroblastoma.

Supplementary methods and figures